54 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.

3 THE IDEAL DIODE EQUATION: DERIVATION GAME PLAN

The II-material has majority carrier electrons that diffuse across the depletion re­ p-bulk region, the depletion region, and the II-bulk region. In Chapter 2 of Volu
gion, are injected into the p-material, and then recombine with majority carrier holes. we have discussed the electrostatic solution for the depletion region in detail. S
Replacement holes are supplied at the metal-p-semiconductor contact by electrons assumptions about the device must be invoked to make the /- VA solution trac
exiting into the external circuit, thereby generating majority carrier holes. The continu­ They are justified since actual devices follow the theory over many decades of cu
ity of current is completed at the metal-n-material contact, which provides the electrons
1. There are no external sources of carrier generation; for example, no light.
for injection.
Figure 3. 5( c), appropriate for reverse biases, illustrates the thermal generation of 2. The depletion approximation and the step junction are applicable.
drift current components near the depletion region edges. For example, the generation 3. The steady-state dc solution is desired; that is, all the dldt terms in the conti
of minority carrier holes in the II-region drifts through the depletion region to the elluations are zero.
p-region, where they become majority carriers and eventually migrate to the metal­ 4. No generation or recombination takes place in the depletion region (what g(
p-material contact. Electrons from the metal annihilate the excess holes, causing an ex­ must come out).
ternal circuit electron to flow. Simultaneously, the thermally generated majority carrier
5. Low-level injection is maintained in the quasi-neutral (bulk) regions of the d
electrons migrate to the metal-n-region contact and into the external circuit, complet­
this means that the number of minority carriers is always much less than the
ing the current loop.
ber of majority carriers in the bulk regions.
The thermally generated electron-hole pairs near the p-side edge of the depletion
region contribute to the reverse current by the electron's drifting across the depletion 6. The electric field for the minority carriers is zero in the bulk regions.
region to the n-material and as a majority carrier moves into the external circuit. At the 7. The bulk regions are uniformly doped; that is, NA and ND are constants.
same time, the thennally generated hole near the p-side depletion region travels to the
With these assumptions the equations of state for the bulk 11- and p-regions r'
p-region contact and is exchanged for an external circuit electron, completing the cur­
to the following minority carrier equations.
rent loop.
It should be pointed out that in Fig. 3. 5(b) the drift current components are not n-type semiconductor:
shown because they are small. The diffusion components in Fig. 3. 5( c) are not shown d 2 fipn fipn
for the same reason. 0= DpjT Tp

3.3 THE IDEAL DIODE EQUATION: DERIVATION GAME PLAN dfipn

J ~ -qDp - d
P x
This section develops a "game plan" for the quantitative solution of the basic semi­
conductor equations as applied to the abrupt p-n junction. The eventual goal is to Pn = PnO + fipn (x)
derive a first-order relationship for the / versus VA dependence of the p-n junction,
p-type semiconductor:
known as the ideal diode equation. Subsequent sections carry out the mathematical
details of the solution outlined in the present section.
The reader is again reminded of the explicit relationship between th;; P-Il junction o = l'~IN d'fi
. n
-_.P - fin I'
diode and the myriad of other solid state devices, and that a thorough understar.ding of dx 2 Tn
the diode is essential to an understanding of other devices. Many of the mathematical
d fill"
derivations presented for the diode are used directly in modeling the bipolar, MOSFET, JN ~ qDv• (X
-1
and junction field effect transistors. Thus, the somewhat excessive time and effort
allotted to the p-n junction will be amply repaid when t;i.alyzing other devices. Also, Ill' = 111'0 + fillp(x)
many of the IC layout design rules for CMOS and other VLSI technologies use these
concepts.
The reader should consult Section 3.4, Volume I, for the details concerning the c
of Eqs. (3.5) to (3.10).
3.3.1 General Considerations The plan of attack is to first solve Eq. (3.5) for fipn(x) in the n-bulk region. Be(
In Chapter 3 of Volume 1, Section 3.4 lists the "elluations of state" for all semiconduc­ Eg. (3.5) is a second-order linear differential elluation, two boundary condition
tor devices: the continuity equations, Poisson's equation, and the current flow equations. needed, one at each end of the region. We will return to a consideration of the boUl
These equations are to be solved in each of the three regions of the p-n junction: the conditions later. Once fipn(x) is obtained, then Eg. (3.6) is used to compute J
56 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.3 THE IDEAL DIODE EQUATION: DERIVATION GAME PLAN

p N
Since the device has only two terminals, the total current through the diode must be a
constant at each point: J(total)

J = constant = IN(x) + Jp(x) (3.11 ) ,,

Therefore, if the minority carrier current density J p (x) is known in the n-bulk region, ~
1:'~~ -]pe x ")
"'E J
the majority carrier current density I N (x) is also known from Eq. (3.11) as (,)

0:.
E
IN(x) = J - Jp(x) (3.12) ~
....,
Arguments for the p-bulk region and the minority carrier electrons are complemen­
tary to those of the minority carrier holes in the n-bulk region. By complementary we %U'lAUVu:
o ~ x
-xI' x"
mean the exchange of p for II and 11 for p. For example, the complement of Eq. (3.5) is
Eq. (3.8). A solution of Eq. (3.8) yields dnp(x) and the use of Eq. (3.9) results in
.. f-Depletion . I •
------ II-bulk region
IN(x), the minority carrier current in the bulk p-region. The majority carrier current is p-bulk region
region
obtained from Eq. (3.11) as
Jp(x"l
Jp(x) = J - IN(x) (3.13)

The perceptive student might ask, "You have outlined a plan of attack for the bulk
rcgions, but what about the current in the depletion region?" The depletion region was
assumed to have no generation or recombination. Therefore, the current through it is a
] - p-bulk Il-bulk --- ]

constant - "what goes in must come out." No carriers are added to or subtracted from
the carrier flux. Figure 3.6 illustrates this point. Note that if the minority carrier dif­ o x"
I N ( -xp)
fusion current is known at the edges of the depletion region, then it is known
throughout the depletion region, and hence at the other edge of the depiction region. Fig. 3.6 Current components at the edges of the depletion region.
Specifically,

JPldepl = Jp(x n ) = dp"

-qDp­ I (3.14) 3.3.2 Boundary Conditions at xn and - Xp
dx X=X I1
In Chapter 2 when VA was applied across the terminals of the diode, the junction
JNldcpl -
_ . _ dill'
I N( -xl') - qDNj
I (3. IS)
tential (Vj) was equated to Vbi - VA under the assumption that the electric field wa
x X=-X p
senti ally zero in the bulk regions. The assumption of "low-level injection" suggests
in the depletion region the additional currents due to the applied voltage are also s
Finally, as illustrated in Fig. 3.6, the total current is simply the sum of Eq. (3.14) and In the depletion region (g 01= 0 and the electron current is the difference between
Eq. (3.15). large current components, JNldrift and JNldiff' We assume that c:g and II have not char
much under low-level injectiorl, that is,
dn
J = Jp(x,,) + I N( - xl') (3.16) I N = qf-t,/lro
(/?
+ qD N-
dx
== 0 c
and solving for the electric field yields Eq. (3.18),
The reader may well ask, "Where does the applied voltage (VA) get into the act? You
started out to derive a J versus VA relationship, but VA has yet to be even mentioned!"
(g = -qDNdll/dx = -DN dn/dx
As it turns out, VA enters via the boundary conditions on Pn(x) and II/X) at the edges of
the depletion region in the solution to the minority carrier diffusion equations.
! qf-t,,1l f-tn II

t
58 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.3 THE IDEAL DIODE EQUATION: DERIVATION GAME PLAN

Applying the Einstein relationship to Eg. (3.18) results in Eg. (3.19), remembering that n,,(x,,) == 11,,0' Since Il ~ /ppo = Ilpo one concludes that

IIp( - X ) 11 eq\'AlkT
~ = -kT dn/dx (3.19)
p pO

q 11
and the injected electron concentration at the boundary - xp is
From the definition of potential, the junction voltage can be written as
xn

Vj = Vbi - VA = - J ~ dJ;; (3.20) tlll (- x p ) = IIpo(eqVAlkf - 1)

-X p

Substituting Eg. (3.19) into Eg. (3.20) yields Note that Egs. (3.28) and (3.29) reduce to their thermal equilibrium values
VA = O.
kT dn/dx kT In(x n.1 Complementary arguments can be used for the hole concentration at the ed
Vj = Vbi - VA = - J X Il
- - - - dx = - In n
q n q n(-'pl
(3.21 )
the depletion regions and result in Eqs. (3.30) and (3.31),
-X p

~. _ V = kT In n,,(x,J (3.22) p,,(x,,) = p"oeqVAlkT

bl A
q np (
-x)
p
Here we have added the extra subscript to help the reader keep track of which bulk
region is being discussed. Remember that for the carriers in the bulk n-region, tll1" == tlp,,(x,,) = p"o(eqvAlkT - 1)
tlpn ~ 11,,0 for the "low-level injection" requirement to be valid. Therefore /l(x,,) = /l"o·
Inverting Eg. (3.22) by cross multiplying and raising both sides to the exponential, we

;SI!FF~~QP'i~J·t~~IiN~~~!~:1
get the electron concentration ratio:
n,,(x,) = eq(Vbi-VAllkT = e[qVbi1kT]e[ -qVAJkT] (3.23)
/lp( - x p)
Long-Base Diode. The final boundary conditions on the excess carrier concen
Solving for Il(-X p ) results in Eq. (3.24), in the P- and n-bulk regions are obtained by assuming that the bulk regions aI"
long, even infinite in length. Since the injected minority carriers have a finite Ii
np(-xp ) = 11,,(x,,)e[-qVbilkTle[qVAikT] (3.24) (Tp and Til)' they cannot survive forever without recombining with a majority
consequent!y,
From Eg. (2.14), for thermal equilibrium,
tlnp(-oo) = 0

~. In[l1no~pol
l
and
= kT (3.25)
bl q ni_ tlp,,( +(0) =0
which can be inverted to obtain Eq. (3.26),
Game Plan Summary

Combining Eg. (3.24) and Eq. (3.26) yields

Il (-x)
e. qVbl.;kT

n (:r ) __,_eqVAlkf
n2
112
_ _i

nnOPpO

. 2
!!i. e qVAikT
(3.26)

(3.27)
I 1. Solve the minority carrier continuity equations in the bulk regions for tlPII(
tlll/X) or p,,(x) and np(x).
2. Apply two boundary conditions to each solution to determine tlp,,(x) and Do,
terms of the applied voltage VA'
3. Determine the currents lAx,,) and IN(-Xp) from the slope of tlnp(x) and tl
P P P -" 11 -xp and x"' respectively, using Eqs. (3.14) and (3.15).
ll"oPpO Ppo
60 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.4 THE IDEAL DIODE EQUATION: DERIVATION

4. The total current is then the sum of the currents at the edges of the depletion region; Dividing by Dp and taking the second term across the equal sign results in Eq. (3
that is, where Lp is defined as the minority carrier diffusion length for holes:
1 = lp(xn) + IN( - xp) (3.34)
d 2f:.Pn(x') = f:.pn(x') = f:.pn(x') (
2 2 '
3.4 THE IDEAL DIODE EQUATION: DERIVATION dx' DpTp Lp

To solve the minority carrier diffusion equations, Eqs. (3.5) and (3.8), we begin by
selecting the special x' and x" coordinate systems as illustrated in Fig. 3.7. The transla­ I Lp = ~DpTp I (cm) C
tion of the working-coordinate system simplifies the analytical form of the solutions and
As discussed in Volume I, this is the average distance a minority carrier hole
avoids unnecessary complications in the analysis.
diffuse before recombining with a majority carrier electron. Equation (3.36) is a
common differential equation and can be solved directly or by Laplace transforms
3.4.1 n-bulk Region, x :2: xn or x' 2: 0' solution is of the form
Equation (3.35) is Eq. (3.5) rewritten in terms of the new variable x' whose origin is at f:.pn(x') = Alex'ILp + A 2e- x'ILp
x = xn .
c:
d 2f:.Pn(x') f:.pn(x') where two boundary conditions are needed to evaluate the constants AI and A 2 . I
Dp dx,2 - -- = 0 (3.35)
Tp Eq. (3.33) of the previous section, the boundary condition at infinity requires

f:.pJXJ) = Ale'" + A 2e- en = Ale'" + 0 = 0 c

11 or p Equation (3.39) can be satisfied only if AI = O. The second boundary condi
Eq. (3.31), requires

f:.Pn(xJ = f:.Pn(O') = pnO(eqVAlkT - 1) = Ale -O'ILp = Al C

p N Therefore the solution for f:.Pn (x) is

f:.Pn (x') = pnO(eqVAlkT - I)e -x'ILp c

or

Pn (x') = PnO + PnO(eqVAlkT - l)e -,'ILp c

,-- - - - PnO
Equation (3.42) is plotted on the right-hand side of Fig. 3.8 for VA > 0 and VA <
I
Il po - - -----I I Following the game plan, we next obtain the hole current by applying Eq. (3.1
I I Ell. (3.41):

-x p oI X
~ x
n
f (x')
.p
= -qD df:.Pn
p dx'
= -qD P (eqVAlkT -
P nO
\) (-I)e-X'ILP
Lp (3
x".. J.. I
0' .. x'
1 p (X ') = qDp PnO (qVAlkT _ 1) -x'ILp (3
e e
Fig. 3.7 Axis selection for bulk regions. Lp
62 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.4 THE IDEAL DIODE EQUATION: DERIVATION

p norp N 3.4.2 p-bulk Region, x1 s; -xp or X" ;::::: 0

~ "nO cxp(qVAlkT) The complementary p-bulk solution to Eq. (3.8) can be established directly or (
1
1 1 • step-by-step in a manner similar to the n-bulk derivation. Simply taking the CI
1
l1
T)
o cxp(qVAlkT) ",
I
1
I i''\.. / l ' n (x')for VA> 0
:\ ment of Egs. (3.41) and (3.42) yields Egs. (3.45) and (3.46),

n,. (x") for V A > 0\ /1

"-I I "'-.
llnp (x") = nl'o(e qVAlkT - l)e -x"ILN

npo _____
. -...,-1
..:.-==-:::._ / : 1---­
II
I ,.'
........
- "no
-----=-­
--­ .
np(x") = npo + Il pO (eqV AlkT - l)e-X"ILN
1.'---1)
• n (")f
X or V A < 0
11 ,. (x") for VA < 0 ~.I, x
-Xp 0 X
n
The minority carrier diffusion length Jar electrons has been defined as
x" .. I
0" I
0' ... X'

Fig.3.8 Minority carrier concentrations in the bulk regions for VA> 0 (-. -. -) and for VA < 0
LN = v'DNTn (cm)
(. .... ) with NA > ND •
This is the averagc distance that the minority carrier electron will diffuse bef(
Evaluating at x = Xn (or x' = 0) yields the hole current in the depletion region, combining with a majority carrier hole. The left-hand side of Fig. 3.8 illustratc~
results for forward and reverse bias. By applying the complement of Eqs. (3.4
D (3.44), the electron currents iN(X") and IN(O") = IN(-xp) are readily determined
Ipldcpl = Ip(x.) = Ip(O') = q~Pno(eqVAlkT - 1) I (3.44)
Lp
IN(x") q ~: Ilpo(eqvAlkT - I)e-x"ILN

Figure 3.9 illustrates the hole current at the edge and throughout the depletion and
n-bulk regions.
D
IN(-X) = q 2 n (eQVAlkT - 1)
p LN pO
p N

}(total)
Figure 3.9 illustrates the electron currcnt components throughout the diode. Rerr

J p (x n ) ."
.,,­ that a current directed along the x"-axis is opposite to the current flow along the
m a negative sign is needed in Eq. (3.15).
~
/' /
~E ~~ /
~
l:! / } 3.4.3 Ideal Diode Equation
'"0.
E The total current is obtained by adding Eg. (3.44) and Eg. (3.49) as outlined in
.s }N (-x p )
...... of the "game plan." The result is called the ideal diode equation or sometin

o _'-=--- -x p
o
i//V/,{//v/d
Xn
....... .. x
Shockley diode equation,

p-bulk region
-I .
Depletion
region
..I" -------­
n-bulk region i = q [ DV
--'- n
L pO
N
+-
D"p
L nO
P
j (eqvAlkT - 1)

Fig.3.9 Current density in the forward-biased case. NA > No·

 64 THE IDEAL DIODE VOLT-AMPERE CHARACTERISTIC 3.5 INTERPRETATION OF RESULTS 6

Multiplying by the area of the junction (A) yields the total current In I

I = lo(eqvAikT - I) (3.51)

where the reverse saturation current has been defined as

10
DNn po + -PliO
= qA [- Dp
LN Lp
J (3.52) In(lo)
I
I

~_[o ·VA ~--------~~ VA

3.5 INTERPRETATION OF RESULTS

(a) (b)
3.5.1 V-/ Relationship Fig. 3.10 1'-11 junction volt-ampere characteristics: (a) linear plot: (b) semilogarithmic plot.
Further examination of the quantitative solution for the ideal diode equation leads to a
deeper insight into the operation of the p-n junction. Figure 3.1O(a) is a linear plot of
Eq. (3.51) illustrating the exponential increase in current with forward bias and the
small, nearly constant, reverse bias current. Examination of the exponential term in
Eq. (3.51) at room temperature, where q/kT = 38.46 V-I, indicates that a forward
bias of 0.10 volts makes the exponential nearly 50 times that of the "one," and there­ 1= qAn;[~ + J!..L] (eqVAlkT - 1) (3.53b
fore, the current is exponentially dependent on the applied voltage. LNNA LpND

I == 10eQVAIkT
It is important to note that ni increases exponentially with temperature and its room
tem'Jerature value is very dependent on the band gap EG of the material. A larger E,
If the natural logarithm is taken then
value results in a smaller value of ni' Also note that the more lightly doped side of th
P-Il junction will produce a larger number of minority calTiers, and thus the larger cur
In(1) = InUo) + qVA rent component. For example, a p+ -n junction has NA ~ No and, from Eq. (3.53),
kT

which is plotted as a straight line in Fig. 3.1O(b). Note the similarity to the form of
10 == qA [ Dp]
LpPIlO = qA
[D p n;]
Lp No (3.5
y = Ax + b when plotted on the semilog axis.
With VA = -0.10 volts, the exponential is about \Iso and very small compared to the
"-I," leaving the current I == -/0 independent of the reverse voltage. The reverse cur­
Siml}arly for a p-n 4 junction,
rent no longer changes, that is, becomes saturated; hence its name "the reverse saturation
current." [See Fig. 3.1O(a).]

n~l
As evidenced by Eq. (3.52) and as visualized in Fig. 3.4, the reverse saturation cur­
rent is determined primarily by the minority carriers flpo thermally generated in the p-bulk 10 == qA [-D N]
npO = qA [DN
-­ (3.5
LN LN NA
region and PlIO generated in the Il-bulk region. Equation (3.52) can be written in terms of
the doping densities N A and No:
Most diodes and many of the P-Il junctions that occur in other devices are of the P +­
10
DN Dp ]
= qA [ LN flpo + Lp PliO
n~+Dp
DN -
qA [- -- fl~J (3.53a)
or n' -P type, and we shall make frequent use of these asymmetrical junctions in futurl
LNNA LpN[) analyses.